Research And Realization Of Key Technology For Fabric Defect Real-Time Detection Based On FPGA

Fabric defect inspection is a vital step for quality control in fabric production. At present,fabric defect inspection in textile industry is primarily performed by manual detection. Thereare drawbacks, such as poor efficiency, low accuracy and higher cost, which were caused byfatigues of human-eye and impacts of individual subjective factors. To improve the efficiencyand reliability of fabric inspection, computer vision technology has been increasingly appliedto detecting fabric defects automatically to replace the hand-counting method. Consideringmutual constraint of the speed and the precision in fabric defect real-time detection system,key technologies Based on FPGA have been studied in this paper. The research subject hassupported by “13115” Major Innovation Project of Shaanxi Province.After investigating application requirement and reading lots of literature at home andabroad, key technologies and a core device have been studied and developed for fabric defectreal-time automation detection based on FPGA by optimizing resources and using parallelcomputing techniques. The main research works are as follows:(1) a technology of externalsynchronous control for line scan CCD camera work;(2) an algorithms and its hardwarerealization technology for detecting fabric defects;(3) a realization technology based on USBfor communicating data between FPGA and PC. Finally, functionality and performance of thecore device are tested on a simulation platform based on machine vision. The result showsthat it reaches expected requirements of function and performance, which provides animportant reference for real-time detection system in textile industry.